An EMC interference source automatic diagnosis system for an electric drive system based on hierarchical analysis of spectral characteristics

Abstract

The present application belongs to the technical field of electromagnetic compatibility test and diagnosis, and solves the problems that the prior art lacks a unified spectrum type discrimination mechanism, it is difficult to consider narrowband harmonic and wideband envelope interference recognition, and the positioning accuracy of a composite interference source is insufficient.The present application includes a spectrum type discrimination module for discriminating spectrum data types, including narrowband harmonic, wideband envelope or composite interference types;a narrowband harmonic diagnosis module;a wideband envelope diagnosis module;a differential common mode auxiliary determination module for determining differential mode or common mode interference dominance;and a result fusion module for weighted calculation based on the confidence and consistency of each result, outputting a candidate interference source and its ranking result.The present application reduces dependence on artificial experience through a spectrum type discrimination mechanism, and improves the adaptability to a composite interference scene through a narrowband harmonic diagnosis module and a wideband envelope diagnosis module to realize simultaneous recognition of multiple interference sources.

Description

Technical Field

[0001] This invention belongs to the field of electromagnetic compatibility testing and diagnosis technology, specifically relating to an automatic diagnostic system for EMC interference sources in electric drive systems based on spectral feature hierarchical analysis. Background Technology

[0002] Electric drive systems are one of the main sources of EMC (Electronic Control Compatibility) issues in new energy vehicles. Current EMC testing and rectification processes typically rely on engineers manually interpreting spectrum diagrams and using their experience to systematically identify and eliminate interference sources.

[0003] In existing technologies, common methods include spectrum identification methods based on human experience and automatic analysis methods based on single-feature rule matching. However, these methods generally suffer from the following problems: a lack of a unified spectrum type discrimination mechanism, making it difficult to distinguish between narrowband harmonics and broadband envelope interference; insufficient adaptability to complex interference scenarios, making it difficult to simultaneously identify multiple interference sources; high dependence on engineering experience, resulting in poor consistency of diagnostic results; and low utilization of test data, making it difficult to form a reusable knowledge system. Summary of the Invention

[0004] To address the problems of existing technologies, such as the lack of a unified spectrum type discrimination mechanism, difficulty in recognizing both narrowband harmonics and broadband envelope interference, and insufficient accuracy in locating composite interference sources, the purpose of this invention is to provide an automatic diagnostic system for EMC interference sources in electric drive systems based on hierarchical analysis of spectrum features.

[0005] To achieve the above-mentioned technical objectives, the technical solution adopted by the present invention is as follows: An automatic diagnostic system for EMC interference sources in electric drive systems based on spectral feature hierarchical analysis, including... The spectrum type discrimination module is used to discriminate the type of spectrum data, including narrowband harmonics, broadband envelope, or composite interference types. Narrowband harmonic diagnostic module; Wideband envelope diagnostic module; Differential-mode or common-mode auxiliary determination module is used to determine the dominance of differential-mode or common-mode interference; The results fusion module performs weighted calculations based on the confidence and consistency of each result, and outputs candidate interference sources and their ranking results.

[0006] This invention uses a spectrum type discrimination module to determine the type of spectrum data. Once the type is determined, a narrowband harmonic diagnosis module and a wideband envelope diagnosis module are used to diagnose the interference source of the corresponding type of spectrum data. If it is a composite interference type, the narrowband harmonic diagnostic module and the wideband envelope diagnostic module are used together for diagnostic analysis. The differential and common mode auxiliary judgment module determines the priority of calling the narrowband harmonic diagnostic module and the wideband envelope diagnostic module when diagnosing composite interference type based on the dominance of differential mode or common mode interference. The results of the diagnosis by the narrowband harmonic diagnostic module and the wideband envelope diagnostic module are then fused by the result fusion module. In the above process, the present invention achieves unified processing of narrowband and broadband interference through a spectrum type discrimination mechanism, reduces reliance on human experience, improves diagnostic consistency and stability, and achieves simultaneous identification of multiple interference sources through narrowband harmonic diagnosis module and broadband envelope diagnosis module, thereby improving adaptability to complex interference scenarios. The functions of each module can be implemented by the processor calling the programs stored in the memory.

[0007] As a preferred technical solution of the present invention, it further includes a data access module and a preprocessing module; The data access module performs unified parsing of various information that requires interference source diagnosis in a multi-source heterogeneous data format, and sends the parsed spectrum data to the preprocessing module. The preprocessing module performs deduplication, sorting, segmentation, standardization, outlier removal, amplitude reshaping, frequency alignment, and noise suppression on the spectrum data. The output of the preprocessing module serves as the input to the spectrum type discrimination module. In the preprocessing module, Segmented standardization normalizes the amplitude based on a preset frequency band division; Frequency alignment is achieved through interpolation or resampling; The preprocessing module outputs a one-dimensional spectrum vector and a two-dimensional energy distribution map.

[0008] In this invention, there are many types of information that need to be diagnosed as interference sources, such as conducted emission test data, radiated emission test data, current method test data, voltage method test data and historical database data. By uniformly analyzing this information, heterogeneous differences can be shielded, joint diagnosis of multi-source data can be achieved, and system maintenance and expansion costs can be reduced. The unified parsed spectrum data needs to be standardized by the preprocessing module to eliminate various non-ideal factors and interference introduced during the data acquisition process. This ensures that the data input to the spectrum type discrimination module has consistency, integrity, and analyzability, thereby improving the accuracy and robustness of subsequent type discrimination and interference diagnosis.

[0009] As a preferred technical solution of the present invention, in terms of spectrum data type identification, a spectrum type discrimination module is included. The dispersion of spectral peaks is quantified by the variance of the peak spacing. Envelope continuity is calculated using the rate of change of amplitude within a sliding window; If the variance of the spectral peak spacing is greater than a preset threshold and the envelope is continuous, the spectral data is determined to be of the composite interference type. If the variance of the spectral peak spacing is greater than the preset threshold and the envelope is not continuous, the spectral data is determined to be narrowband harmonics. If the variance of the spectral peak spacing is less than a preset threshold and the envelope is continuous, then the spectral data is determined to be a broadband envelope. If the variance of the spectral peak spacing is less than the preset threshold and the envelope is not continuous, then the spectral data is determined to have no effective interference. The spectrum type discrimination module determines whether the spectrum data has broadband characteristics by calculating the ratio of the spectrum data bandwidth to the receiver resolution bandwidth, thus avoiding misjudgment.

[0010] As a preferred technical solution of the present invention, the retrieval methods performed by the narrowband harmonic diagnosis module include fundamental frequency retrieval, harmonic interval retrieval, full-band blind retrieval and single-frequency point blind retrieval. Before the narrowband harmonic diagnosis module performs retrieval, the spectrum data undergoes frequency offset processing, amplitude envelope removal, noise threshold screening and harmonic extraction. Full-band blind search performs peak detection on spectrum data to form a candidate peak set. Using the candidate peak frequency points as input and the distance between each peak frequency point as matrix elements, a distance matrix is ​​constructed to detect whether there is a peak sequence that repeats at equal intervals. When a preset number of consecutive peaks at equal intervals are detected, they are determined to be the same harmonic group. Multiple harmonic groups form a harmonic group library. Single-frequency blind search retrieves harmonic groups containing the target frequency from the harmonic group library and outputs the fundamental frequency, harmonic spacing, frequency band range, number of peaks, and corresponding candidate interference source labels of the harmonic group. Fundamental frequency retrieval utilizes known fundamental frequencies and characteristic parameters of potential interference sources. Based on the rule that harmonic frequencies are integer multiples of the fundamental frequency, it generates a set of theoretical harmonic frequency points, thereby matching the corresponding harmonic sequences in the spectral data. Harmonic interval retrieval is based on frequency multiples. The frequency interval between adjacent anchor points in the same harmonic sequence is used as the step size. A bidirectional equal step size retrieval can be performed within the target frequency band to obtain the interval.

[0011] As a preferred embodiment of the present invention, the broadband envelope diagnostic module includes a feature weighted fusion unit and extracts features using a spectrum-image dual-modal fusion model. In the spectrum-image dual-modal fusion model, The first feature extraction network is used to extract spectral sequence features; The second feature extraction network is used to extract features from the two-dimensional energy map; The feature weighted fusion unit is used to perform weighted fusion of two types of features and output the multi-label interference source identification result.

[0012] This invention combines the ability of one-dimensional spectrum to faithfully represent the overall structure of continuous envelope with the ability of two-dimensional energy map to enhance the representation of local energy distribution and image-like detail features. By adaptively fusing the two types of features, the ability to capture key envelope features and the anti-interference performance can be improved in scenarios with complex backgrounds, broadband noise disturbances, narrowband spike interference, and details differences.

[0013] As a preferred technical solution of the present invention, the differential-common mode auxiliary determination module calculates the difference in amplitude and the trend of spectrum change of the positive and negative poles based on the single-terminal voltage data of the high voltage or low voltage positive and negative poles, and determines the dominance of differential-mode or common-mode interference.

[0014] As a preferred technical solution of the present invention, it also includes a rectification support module, a report output module, and a database unit; The rectification support module is used to generate rectification suggestions based on the type of interference source and the dominant coupling mode, and establishes a mapping relationship between the rectification suggestions and the interference type. The report output module receives rectification suggestions from the rectification support module and uses them to generate a diagnostic report. The diagnostic report includes the following information: out-of-limit frequency information, interference type, candidate interference source, harmonic structure information, identification probability, and rectification suggestions. The database unit stores the output information of the result fusion unit and the rectification support module, improving the efficiency of test data utilization and supporting knowledge accumulation and reuse.

[0015] The beneficial effects of this invention are as follows: By using a spectrum type discrimination mechanism, it achieves unified processing of narrowband and broadband interference, reduces reliance on human experience, and improves diagnostic consistency and stability. By using a narrowband harmonic diagnosis module and a broadband envelope diagnosis module, it achieves simultaneous identification of multiple interference sources, thereby improving adaptability to complex interference scenarios. Attached Figure Description

[0016] The present invention can be further illustrated by the non-limiting embodiments given in the accompanying drawings; Figure 1 This is a block diagram of the overall architecture of the present invention; Figure 2 This is a flowchart of the data access module and preprocessing module of the present invention; Figure 3 This is a flowchart illustrating the spectrum type discrimination method of the present invention. Figure 4 This is a schematic diagram of the narrowband harmonics and broadband envelope characteristics in this invention. The upper part represents the narrowband harmonics, and the lower part represents the broadband envelope. Figure 5 This is a flowchart illustrating the workflow of the narrowband harmonic diagnosis module in this invention. Figure 6 This is a flowchart illustrating the workflow of the broadband envelope diagnostic module in this invention. Figure 7 This is a flowchart of the rectification support module and the report output module in this invention. Detailed Implementation

[0017] The technical solutions of the present invention will be described in detail below with reference to specific embodiments and accompanying drawings. The embodiments described herein are specific implementations of the present invention, used to illustrate the concept of the present invention; these descriptions are explanatory and exemplary, and should not be construed as limiting the implementation methods or the scope of protection of the present invention. In addition to the embodiments described herein, those skilled in the art can employ other obvious technical solutions based on the content disclosed in the claims and specification of this application. These technical solutions include those that make any obvious substitutions and modifications to the embodiments described herein. Example

[0018] like Figures 1-7 As shown, this embodiment provides an automatic diagnostic system for EMC interference sources in an electric drive system based on spectral feature hierarchical analysis. The system is deployed on a host computer with an EMC test bench. After the test, the data access module automatically reads the frequency domain data file of conducted or radiated emissions and imports prior information such as operating conditions, test methods, and known interference source fundamental frequencies according to the test plan. After the preprocessing module completes the spectrum standardization, the spectrum type discrimination module performs stream splitting based on the characteristics of discrete peaks and continuous envelopes; If narrowband harmonic features are identified, the narrowband harmonic diagnosis module calls full-band blind search and single-frequency point blind search to output multiple candidate harmonic groups; if wideband envelope features are identified, the wideband envelope diagnosis module calls the spectrum-image dual-mode fusion model to output the probability of composite interference sources; if the system simultaneously receives high and low voltage single-ended voltage data, the differential common-mode auxiliary determination module synchronously provides the dominant noise mode. Based on this, the result fusion module outputs the top-ranked candidate interference sources and submits the results to the rectification support module to generate rectification suggestions. Finally, the report output module generates a diagnostic report containing information on out-of-limit frequency points, interference types, candidate interference sources, harmonic structure information, and identification probabilities.

[0019] This embodiment achieves unified processing of narrowband and broadband interference through a spectrum type discrimination mechanism, reducing reliance on human experience and improving diagnostic consistency and stability. It also enables simultaneous identification of multiple interference sources through narrowband harmonic diagnosis modules and broadband envelope diagnosis modules, thereby improving adaptability to complex interference scenarios.

[0020] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the present invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. An automatic diagnostic system for EMC interference sources in an electric drive system based on hierarchical analysis of spectral features, characterized in that: include The spectrum type discrimination module is used to discriminate the type of spectrum data, including narrowband harmonics, broadband envelope, or composite interference types. Narrowband harmonic diagnostic module; Wideband envelope diagnostic module; Differential-mode or common-mode auxiliary determination module is used to determine the dominance of differential-mode or common-mode interference; The results fusion module performs weighted calculations based on the confidence and consistency of each result, and outputs candidate interference sources and their ranking results.

2. The automatic diagnostic system for EMC interference sources of an electric drive system based on spectral feature hierarchical analysis according to claim 1, characterized in that: It also includes a data access module and a preprocessing module; The data access module performs unified parsing of various information that requires interference source diagnosis in a multi-source heterogeneous data format, and sends the parsed spectrum data to the preprocessing module. The preprocessing module performs deduplication, sorting, segmentation, standardization, outlier removal, amplitude reshaping, frequency alignment, and noise suppression on the spectral data. The output of the preprocessing module serves as the input to the spectral type discrimination module.

3. The automatic diagnostic system for EMC interference sources in an electric drive system based on spectral feature hierarchical analysis according to claim 2, characterized in that: In the preprocessing module, Segmented standardization normalizes the amplitude based on a preset frequency band division; Frequency alignment is achieved through interpolation or resampling; The preprocessing module outputs a one-dimensional spectrum vector and a two-dimensional energy distribution map.

4. The automatic diagnostic system for EMC interference sources in an electric drive system based on spectral feature hierarchical analysis according to claim 1, characterized in that: Spectrum type identification module, The dispersion of spectral peaks is quantified by the variance of the peak spacing. Envelope continuity is calculated using the rate of change of amplitude within a sliding window; If the variance of the spectral peak spacing is greater than a preset threshold and the envelope is continuous, the spectral data is determined to be of the composite interference type. If the variance of the spectral peak spacing is greater than the preset threshold and the envelope is not continuous, the spectral data is determined to be narrowband harmonics. If the variance of the spectral peak spacing is less than a preset threshold and the envelope is continuous, then the spectral data is determined to be a broadband envelope.

5. The automatic diagnostic system for EMC interference sources of an electric drive system based on spectral feature hierarchical analysis according to claim 4, characterized in that: The spectrum type discrimination module determines whether the spectrum data has broadband characteristics by calculating the ratio of the spectrum data bandwidth to the receiver resolution bandwidth.

6. The automatic diagnostic system for EMC interference sources of an electric drive system based on spectral feature hierarchical analysis according to claim 1, characterized in that: The narrowband harmonic diagnostic module performs two search methods: full-band blind search and single-frequency point blind search. Full-band blind search performs peak detection on spectrum data to form a candidate peak set. Using the candidate peak frequency points as input and the distance between each peak frequency point as matrix elements, a distance matrix is ​​constructed to detect whether there is a peak sequence that repeats at equal intervals. When a preset number of consecutive peaks at equal intervals are detected, they are determined to be the same harmonic group. Multiple harmonic groups form a harmonic group library. Single-frequency blind search retrieves harmonic groups containing the target frequency from the harmonic group library and outputs the fundamental frequency, harmonic spacing, frequency band range, number of peaks, and corresponding candidate interference source labels of the harmonic group.

7. The automatic diagnostic system for EMC interference sources of an electric drive system based on spectral feature hierarchical analysis according to claim 6, characterized in that: The narrowband harmonic diagnostic module also performs fundamental frequency retrieval and harmonic interval retrieval.

8. The automatic diagnostic system for EMC interference sources of an electric drive system based on spectral feature hierarchical analysis according to claim 1, characterized in that: The broadband envelope diagnostic module includes a feature weighting fusion unit and uses a spectrum-image dual-modal fusion model to extract features. In the spectrum-image dual-modal fusion model, The first feature extraction network is used to extract spectral sequence features; The second feature extraction network is used to extract features from the two-dimensional energy map; The feature weighted fusion unit is used to perform weighted fusion of two types of features and output the multi-label interference source identification result.

9. The automatic diagnostic system for EMC interference sources of an electric drive system based on hierarchical analysis of spectral features according to claim 1, characterized in that: The differential-mode and common-mode auxiliary judgment module calculates the amplitude difference and spectrum change trend of the positive and negative poles based on the single-terminal voltage data of the high-voltage or low-voltage positive and negative poles, and determines the dominance of differential-mode or common-mode interference.

10. The automatic diagnostic system for EMC interference sources of an electric drive system based on hierarchical analysis of spectral features according to claim 1, characterized in that: It also includes a rectification support module, a report output module, and a database unit; The rectification support module is used to generate rectification suggestions based on the type of interference source and the dominant coupling mode. The report output module receives rectification suggestions from the rectification support module and uses them to generate a diagnostic report. The database unit stores the output information of the result fusion unit and the rectification support module.

Images